From: Brian P. <br...@va...> - 2000-11-28 17:38:39
|
Gareth Hughes wrote: > > Brian Paul wrote: > > > > Gareth Hughes wrote: > > > > > > Is anyone else getting screwed-up texture coords on the cylinder? I've > > > seen this with hw acceleration and LIBGL_ALWAYS_INDIRECT=1 on both the > > > trunk and ati-4-1-1 branch (which has identical Mesa code to the trunk > > > now). > > > > No, I seems OK to me. err--> "it" > > Hmm, okay. I'll have another look, make sure I haven't screwed up the > build/install somehow... > > > The problem is in the auto-normal generation calculation for > > evaluator patches. At the very top and bottom of the teapot > > a number of control points collapse on top of each other so the > > partial deriviates, err----> "derivatives" > and hence their cross product normal, are > > <0,0,0>. There's probably some trick to fixing that but I've > > never looked into it. > > I'll stick it waaaay down the end of my todo list then, shall I? ;-) Sure. It's been near the bottom of my list for a few years now. :) -Brian |